The research program is designed to investigate complex repertoires of behavior. One set of experiments would study basic variables determining the repeated acquisition of long response sequences, using as a dependent variable the maximum sequence that the subject can learn. Several stimulus fading techniques would be explored along with a study of reinforcement frequency. The general issue is whether a major factor in the effectiveness of fading is the maintenance of an adequate frequency of reinforcement. A second set of experiments would investigate other variables, using the learning of long sequences as a behavioral baseline. Also proposed is a parametric study of delay of punishment and of delay resetability, using first "errors" and then "corrects" as the punished response. The broader issue concerns the amount of generalized suppression spreading from the punished behavior to other behavior. Certain drugs will be examined for possible performance enhancement and certain toxic agents for possible behavioral toxicity. A third set of experiments would investigate matching-to-sample repertoires with several techniques for generating quantifiably difficult discriminations.